fullofcaffeine

[fullofcaffeine @FullOfCaffeine]

@levelsio What about CTT in Portugal?

[Mitchell Hashimoto @mitchellh]

We've gone really quickly from "local models are dogshit" to "local models are good actually" (like, a 12 month window from A to B). I don't think they're actually good ENOUGH yet. We need an Opus 4.5 quality local model. When that happens, I think the world will spill over.

Opus 4.5 is/was amazing, and is more than good enough for almost all tasks still as long as you pair with a frontier-level planner/judge.

It'll still require a hugely expensive machine to run it, I'm sure, like a $5K or more laptop or mac studio. But, that's going to be pennies compared to the API costs plus all the benefits of guaranteed privacy and so on.

[fullofcaffeine @FullOfCaffeine]

@arvidkahl @levelsio Yes! I used Wise to transfer lump sums before and it used to work great. I understand why they are going through that in the EU, but they should be more transaprent and try not to fuck up the UX too much, or they'll lose a lot of customers.

[fullofcaffeine @FullOfCaffeine]

@markerdmann @levelsio IBKR is great for that, indeed.

[fullofcaffeine @FullOfCaffeine]

@eurofounder @levelsio You forgot the /s at the end I guess? :)

[fullofcaffeine @FullOfCaffeine]

@spartacus_42069 @eurofounder @levelsio Shady South American country? If you mean Brazil, regulations are on par wtih Europe, if not worse (or better, depending on how you view it).

[fullofcaffeine @FullOfCaffeine]

@bobzhang1988 @deepseek_ai How does it compare to GPT 5.5?

[fullofcaffeine @FullOfCaffeine]

@J_dev363 @rezoundous I think he meant limits are draining too fast.

[fullofcaffeine @FullOfCaffeine]

@Jaimieborodo @mitsuhiko And runtime(s) :)

[fullofcaffeine @FullOfCaffeine]

@VictorTaelin What thinking level?

[fullofcaffeine @FullOfCaffeine]

@burkov @thsottiaux Yes! It requires quite a bit of steering to produce the output that Opus often gets right the first time. It means that Opus is often much better for documentation.,

[fullofcaffeine @FullOfCaffeine]

GPT 5.4 is not good at writing user-friendly docs. To be more specific: GPT usually tends to be way too terse and technical and use terms that might not be familiar to the end-user, e.g in a public README. Steering often fixes that though. Opus 4.5 is much better at capturing my intentions given the context, it's prose is often much better by default.

And ofc, its design skills. Opus is still better!

And finally - and foremost - the current Codex outage, happening just now: https://t.co/3oLfNGomVI.

[fullofcaffeine @FullOfCaffeine]

@RCallsign @sudoingX And I mean the commercial aspect, they literally have us by the b*lls if they want.

[fullofcaffeine @FullOfCaffeine]

@RCallsign @sudoingX Well, I can only hope it will become cost and time effective! I use and like the closed-weight frontier models, but do you really want to live in a world where you depend solely on them? That's pretty depressing and dangerous.

[fullofcaffeine @FullOfCaffeine]

@sudoingX I have a sparing Quadro RTX 5000 with 16GB. Wondering if it'd at least get close to that? Great content, btw!

[fullofcaffeine @FullOfCaffeine]

@rezoundous Yes!

[fullofcaffeine @FullOfCaffeine]

You can do that, and it might work reasonably well. The difference is that the agent needs to do less and the oracle implements it deterministically. And oracle implements and abstracts the browser communication/orchestration side (if you're using the browser) which would be a nightmare for the agent to do everytime. If you're using the API only, you might be able to implement a similar system via instructions via AGENTS.md or a skill or just prompt it everytime.

[fullofcaffeine @FullOfCaffeine]

@nate_yiu @VictorTaelin Would like to know as well!

[fullofcaffeine @FullOfCaffeine]

@VictorTaelin You can use something like https://t.co/R4DjmY98yk to automate that. The browser integration is flaky, but I've made it work better locally, with some additional fixes. It also supports Pro via the API, which is, of course, reliable.

[fullofcaffeine @FullOfCaffeine]

@VictorTaelin Pro is a beast. I love it. This workflow you described is great, it's almost a "secret sauce"-kind-of-thing. Too bad it's very slow, but it's worth it most of the time.